Yield comparison of push and pull control methods on production systems with unreliable machines

In just-in-time manufacturing, the pull control method is used to coordinate feeding of work centers. It is reported to be responsible for the dramatic reduction of work-in-process inventory, which results in quick detection and correction of machine malfunctioning. As a consequence, production yield is increased and throughput time is decreased simultaneously. In this paper we analytically verify this premise and provide insight on its validity by using a tractable theoretical model of a simple production system. We show that, for a production system in which some machines may malfunction and produce defective parts, the just-in-time pull control method achieves a higher yield than the traditional push control method with equal production throughput.